Fuel flow abnormality identification method, device and electronic equipment

Abstract

The present invention provides a fuel flow anomaly identification method, device and electronic equipment, wherein the method comprises: obtaining flow data of a ship at sea; wherein the flow data comprises an inlet flow rate and an outlet flow rate of fuel of a diesel engine of the ship; inputting the flow data into a pre-trained deviation fault identification model, reconstructing the flow data through the deviation fault identification model, and outputting reconstructed data corresponding to the flow data; calculating a deviation value between the flow data and the reconstructed data; if the deviation value exceeds a pre-set deviation threshold, it is determined that a deviation fault occurs in the diesel engine of the ship, and there is no need to establish a complex mechanism model or an expert knowledge base, and the deviation fault of the flow meter can be identified quickly and accurately, which is simple and effective, has a high identification efficiency, and can monitor the deviation fault of the flow meter online in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of flow monitoring, and in particular to a method, a device and an electronic device for identifying abnormal fuel flow. Background Art

[0002] When a ship is sailing, the ship's diesel engine runs under a certain load. The diesel engine needs to continuously consume a certain amount of fuel to output power to maintain the ship's navigation. At this time, the diesel engine flow meter records the diesel engine fuel inlet and outlet flow. The diesel engine flow meter often has deviation failures during continuous operation. Since the diesel engine continuously consumes a certain amount of fuel, when the ship is sailing, the diesel engine flow meter fuel inlet rate is usually greater than the outlet rate, so it is difficult to identify the flow meter deviation failure.

[0003] The commonly used method is to compare the deviation of the flow meter inlet and outlet rates when the diesel engine load is zero when the ship is stopped. When the deviation value exceeds the set value, the flow meter has a deviation failure. However, this method has obvious defects. The time of ship stoppage is uncertain. If the diesel engine flow meter has a deviation failure when the ship is sailing, it will seriously affect the calculation of diesel engine fuel consumption, causing unnecessary misjudgment to the crew and causing economic losses. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a fuel flow abnormality identification method, device and electronic equipment, which avoids relying on the accuracy of the mechanism model and can quickly and accurately identify the deviation fault of the flow meter by only using the historical samples of the ship diesel flow meter.

[0005] In a first aspect, an embodiment of the present invention provides a method for identifying fuel flow anomalies, wherein the method comprises: obtaining flow data of a ship at sea; wherein the flow data comprises an inlet flow rate and an outlet flow rate of fuel of a diesel engine of the ship; inputting the flow data into a pre-trained deviation fault identification model, reconstructing the flow data through the deviation fault identification model, and outputting reconstructed data corresponding to the flow data; calculating a deviation value between the flow data and the reconstructed data; if the deviation value exceeds a pre-set deviation threshold, determining that a deviation fault occurs in the diesel engine of the ship.

[0006] Further, the deviation fault identification model includes an encoder, a hidden layer representation layer and a decoder connected in sequence.

[0007] Furthermore, the method also includes: initializing a pre-established deviation fault identification model; obtaining a pre-constructed training sample, and training the deviation fault identification model according to the following iterative process until a preset training stop condition is met: inputting the training sample into the encoder of the initialized deviation fault identification model, and mapping the training sample to the latent space corresponding to the hidden layer representation layer; reconstructing the training sample through the decoder, and outputting the reconstructed sample corresponding to the training sample; calculating the loss function of the training sample and the reconstructed sample; wherein the training stop condition includes at least one of the following conditions: the number of iterative processes reaches a preset number of times, or the loss function converges.

[0008] Furthermore, the method also includes: obtaining historical flow data of the ship during navigation, the historical flow data including the historical fuel inlet flow rate and the historical fuel outlet flow rate of the ship's diesel engine; preprocessing the historical flow data to obtain training samples, wherein the preprocessing process includes at least data cleaning and data normalization.

[0009] Furthermore, the method further includes: calculating residual values ​​of the training sample and the reconstructed sample; performing a kernel density analysis on the residual values ​​to find a deviation threshold for characterizing the normality and difference of the training sample.

[0010] Furthermore, the method includes: the method also includes: mapping the training sample to the potential space corresponding to the hidden layer representation layer using the following formula:

[0011] h=f(wx+b)

[0012] Among them, f(wx+b) is the coding transfer function, x is the training sample, w is the coding weight of the coding transfer function, h is the eigenvalue of the hidden layer representation layer, and b is the bias parameter of the coding transfer function.

[0013] Furthermore, the method includes: the decoder reconstructs the training sample using the following formula, and outputs the reconstructed sample corresponding to the training sample:

[0014] x'=g(vh+b')

[0015] Among them, g(vh+b') is the decoding transfer function, x' is the reconstructed sample, v is the decoding weight of the decoding transfer function, h is the eigenvalue of the hidden layer representation layer, and b' is the bias parameter of the decoding transfer function.

[0016] Furthermore, the loss function is expressed as the following formula:

[0017] L=||x-x'|| 2

[0018] Among them, x is the training sample, x' is the reconstructed sample, and L is the loss function.

[0019] In a second aspect, an embodiment of the present invention provides a fuel flow anomaly identification device, wherein the device includes: an acquisition module, used to acquire flow data of a ship at sea; wherein the flow data includes an inlet flow rate and an outlet flow rate of the ship's diesel engine fuel; a training module, used to input the flow data into a pre-trained deviation fault identification model, reconstruct the flow data through the deviation fault identification model, and output reconstructed data corresponding to the flow data; a calculation module, used to calculate the deviation value between the flow data and the reconstructed data; a determination module, used to determine that a deviation fault occurs in the ship's diesel engine if the deviation value exceeds a pre-set deviation threshold.

[0020] In a third aspect, an embodiment of the invention provides an electronic device, which includes a processor and a memory, the memory stores computer-executable instructions that can be executed by the processor, and the processor executes the computer-executable instructions to implement any of the above methods.

[0021] Other features and advantages of the present disclosure will be set forth in the following description, or some features and advantages may be inferred or unambiguously determined from the description, or may be learned by implementing the above-mentioned technology of the present disclosure.

[0022] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0024] Figure 1 A flowchart of a method for identifying abnormal fuel flow provided by an embodiment of the present invention;

[0025] Figure 2 A flowchart of another method for identifying abnormal fuel flow provided by an embodiment of the present invention;

[0026] Figure 3 A structural diagram of a deviation fault identification model provided by an embodiment of the present invention;

[0027] Figure 4 A schematic diagram of the structure of a fuel flow abnormality identification device provided by an embodiment of the present invention;

[0028] Figure 5A structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] At present, the diesel engine flow meter can record the diesel engine fuel inlet and outlet flow. The diesel engine flow meter often has deviation faults during continuous operation. Since the diesel engine continuously consumes a fixed amount of fuel, when the ship is sailing, the diesel engine flow meter fuel inlet rate is usually greater than the outlet rate. Therefore, it is difficult to identify the flow meter deviation fault. The commonly used method is to compare the deviation of the flow meter inlet and outlet rates when the diesel engine load is zero when the ship is stopped. When the deviation value exceeds the set value, the flow meter has a deviation fault. Based on this, a fuel flow anomaly identification method, device and electronic device provided in an embodiment of the present invention can use the historical normal data of the ship diesel flow meter to train the automatic encoder, reconstruct the normal data sample, calculate the residual, and use the kernel density estimation and set the deviation fault threshold. The present application can identify the diesel engine deviation fault online. When the deviation obtained by inputting the normal flow data into the deviation fault identification model exceeds the preset threshold, it can be determined that the ship's diesel engine has a deviation fault.

[0031] To facilitate understanding of this embodiment, a fuel flow abnormality identification method disclosed in an embodiment of the present invention is first introduced in detail.

[0032] The embodiment of the present invention provides a method for identifying abnormal fuel flow. Figure 1 A flowchart of a method for identifying abnormal fuel flow provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method comprises the following steps:

[0033] Step S101, obtaining the flow data of the ship underway; wherein the flow data includes the inlet flow rate and outlet flow rate of the diesel engine fuel of the ship;

[0034] Specifically, the flow data of a ship underway can be obtained by using a diesel engine flow meter to record the diesel engine fuel inlet or outlet flow. The diesel engine fuel inlet flow rate and outlet flow rate of the ship can be obtained without stopping the ship. The obtained flow data can be encoded by an automatic encoder, and the original flow data can be cleaned and normalized to obtain flow data that can be recognized by the following deviation fault recognition model.

[0035] Step S103, inputting the flow data into a pre-trained deviation fault recognition model, reconstructing the flow data through the deviation fault recognition model, and outputting reconstructed data corresponding to the flow data;

[0036] Specifically, in order to process the data, the data can often be reconstructed by inputting it into a trained model. Therefore, the above-mentioned traffic data can be processed by the above-mentioned trained deviation fault identification model, and the corresponding reconstructed data can be output to facilitate the subsequent calculation of the deviation value.

[0037] Step S105, calculating the deviation value between the flow data and the reconstructed data;

[0038] In practical applications, the difference between the calculated flow data and the reconstructed data can be used to obtain the above-mentioned deviation value, that is, the inlet and outlet flow rates of the ship's diesel engine fuel obtained by actual measurements and the theoretical difference between the reconstructed data that should be obtained through the above-mentioned deviation fault identification model can be used to obtain the inlet and outlet flow deviation value of the ship's diesel engine.

[0039] Step S107: If the deviation value exceeds a preset deviation threshold, it is determined that a deviation fault occurs in the diesel engine of the ship.

[0040] In actual applications, staff can determine the above-mentioned deviation threshold based on historical data or using the parameters provided by the fuel meter manufacturer, and compare the inlet and outlet flow deviation value of the diesel engine obtained above with the preset threshold. If the above-mentioned deviation value is greater than the flow threshold, it can be determined that the diesel engine has a deviation fault.

[0041] A fuel flow anomaly identification method provided in an embodiment of the present invention can be used for online monitoring of deviation faults of fuel flow meters of in-flight diesel engines of ships. The principle is to use the historical normal data of the ship's diesel flow meter to train a deviation fault identification model, reconstruct normal data samples, calculate the residual of the normal flow data and its model reconstructed data, use the kernel density estimation method to estimate the parameters based on historical data or provided by the fuel meter manufacturer to set the flow meter deviation fault threshold, calculate the residual of the online real-time data, and when the deviation value obtained by inputting the flow data into the model exceeds the preset threshold, it can be determined that the diesel engine has a deviation fault. Since previous solutions often use expert knowledge base methods, the method provided in the embodiment of the present application avoids the use of expert database methods, but uses the historical normal data of the ship's diesel flow meter to train the model. There is no need to establish a complex mechanism model or an expert knowledge base, and the deviation fault of the flow meter can be quickly and accurately identified. It is simple and effective, has high recognition efficiency, and can monitor the deviation fault of the flow meter in real time online.

[0042] Figure 2 A flowchart of another method for identifying abnormal fuel flow is shown. Figure 1 The method is implemented based on a fuel flow anomaly identification method shown in the figure, which further explains the process of training the deviation fault identification model. Figure 3 The structure diagram of a deviation fault recognition model shown in FIG. 1 includes an encoder, a hidden layer representation, and a decoder. Figure 2 As shown, the method comprises the following steps:

[0043] Step S201, initializing a pre-established deviation fault identification model;

[0044] In order to train the above-mentioned deviation fault identification model, the initialization weights in the above-mentioned deviation fault identification model, such as the encoding weights of the encoding transfer function or the encoding weights of the decoding transfer function in the deviation fault identification model, are initialized before the training operation, so as to perform subsequent training operations.

[0045] In the embodiment of the present application, the fuel flow anomaly identification includes both the inlet flow rate and the outlet flow rate of the ship's diesel engine fuel. Therefore, the deviation thresholds of the outlet flow speed and the inlet flow rate can be set separately. For the inlet flow and the outlet flow, this method can monitor both at the same time or monitor either one of them, and no specific limitation is made here.

[0046] Step S203, obtaining a pre-constructed training sample, and training the deviation fault recognition model according to the following iterative process until a preset training stop condition is met;

[0047] In order to make the flow data information recognizable by the deviation fault recognition model, the encoder in the above-mentioned deviation fault recognition model can be used to select the historical normal flow data of the diesel engine, and the obtained historical normal flow data can be used to train the deviation fault recognition model.

[0048] Step S205, input the training sample to the encoder of the initialized deviation fault recognition model, and map the training sample to the latent space corresponding to the hidden layer representation layer; reconstruct the training sample through the decoder, and output the reconstructed sample corresponding to the training sample; calculate the loss function of the training sample and the reconstructed sample;

[0049] The training stop condition includes at least one of the following conditions: the number of iterations reaches a preset number, or the loss function converges.

[0050] In practical applications, when the stop conditions of the above-mentioned training deviation fault identification model are met, it is proved that the deviation fault identification model can be put into practical application, and the traffic data in the practical embodiment of the present application can be reconstructed and subsequent steps can be executed.

[0051] Specifically, the following formula can be used to map the training samples to the potential space corresponding to the hidden layer representation layer:

[0052] h=f(wx+b)

[0053] Among them, f(wx+b) is the coding transfer function, x is the training sample, w is the coding weight of the coding transfer function, h is the eigenvalue of the hidden layer representation layer, and b is the bias parameter of the coding transfer function.

[0054] In practical applications, the normal inlet flow rate and outlet flow rate can be set to x1 and x2 respectively as training samples of the above encoder, and the corresponding hidden layer feature values ​​h1 and h2 can be obtained for x1 and x2 respectively.

[0055] Specifically, the decoder reconstructs the training sample using the following formula and outputs the reconstructed sample corresponding to the training sample:

[0056] x'=g(vh+b')

[0057] Among them, g(vh+b') is the decoding transfer function, x' is the reconstructed sample, v is the decoding weight of the decoding transfer function, h is the eigenvalue of the hidden layer representation layer, and b' is the bias parameter of the decoding transfer function.

[0058] In practical applications, the corresponding h1 and h2 obtained by using the above-mentioned inlet flow rate and outlet flow rate are input into the above-mentioned decoding transfer function to obtain the reconstructed data x′1 and x′2 of the inlet flow rate and outlet flow rate respectively.

[0059] Specifically, the loss function is expressed as the following formula:

[0060] L=||x-x'|| 2

[0061] Among them, x is the training sample, x' is the reconstructed sample, and L is the loss function.

[0062] As judged above, when the loss function converges, one of the training stop conditions can be achieved.

[0063] Step S207, obtaining historical flow data of the ship during navigation, the historical flow data including the historical fuel inlet flow rate and the historical fuel outlet flow rate of the diesel engine of the ship;

[0064] In practical applications, the historical flow data of the ship underway can be used to organize the data, thereby obtaining the historical fuel inlet flow rate and the historical fuel outlet flow rate of the above-mentioned diesel engine during normal operation. Therefore, the historical data can be used as a reference to determine whether the flow data of the ship underway is abnormal.

[0065] Step S209, preprocessing the historical traffic data to obtain the training samples, wherein the preprocessing process at least includes data cleaning and data normalization processing;

[0066] Since the historical traffic data of ships at sea are complex and diverse, there will be certain duplicate information or errors for the overall sample. Therefore, data cleaning technology can be used to review and verify the historical traffic data in order to delete duplicate information, correct existing errors, and provide data consistency.

[0067] Due to the characteristics of the encoder, the historical traffic data needs to be converted from a dimensional expression to a dimensionless expression. Therefore, the above historical traffic data can be converted from a number to a decimal between (0, 1). This is mainly proposed for the convenience of data processing. Mapping the data to the range of 0 to 1 for processing can make it more convenient and quick to process the sample data in the above model.

[0068] Step S211, calculating the residual value of the training sample and the reconstructed sample;

[0069] In practical applications, the normal import flow rate and export flow rate can be set to x1 and x2 and the reconstructed data x′1 and x′2 of the import flow rate and export flow rate of the training samples of the above encoder can be used for difference calculation to obtain the residual value res of the training sample and the reconstructed sample.

[0070] Step S213, performing a kernel density analysis on the residual value to find a deviation threshold for characterizing the normality and difference of the training sample.

[0071] Specifically, by performing kernel density analysis on the residual value, a deviation threshold δ for characterizing the normality and difference of the training sample can be found. Correspondingly, the deviation thresholds of the inlet flow rate and the outlet flow rate are δ1 and δ2, respectively.

[0072] Among them, kernel density estimation is used in probability theory to estimate the unknown density function, which is one of the non-parametric test methods. Since the above-mentioned deviation threshold δ is large in number and there may be different values ​​of deviation threshold δ for ships under normal operation, the kernel density estimation method can be used to estimate the deviation threshold, so as to obtain data that meets the actual requirements.

[0073] Step S215, obtaining the flow data of the ship underway; wherein the flow data includes the inlet flow rate and outlet flow rate of the diesel engine fuel of the ship;

[0074] In practical applications, if the above-obtained inlet flow rate and outlet flow rate are set as y1 and y2 as real-time data, the real-time data can be directly reconstructed.

[0075] Step S217, inputting the flow data into a pre-trained deviation fault identification model, reconstructing the flow data through the deviation fault identification model, and outputting reconstructed data corresponding to the flow data;

[0076] In practical applications, based on the above process of reconstructing the flow data using the deviation fault identification model, the above inlet flow rate and outlet flow rate are set to y1 and y2, and the corresponding reconstructed data are y′1 and y′2.

[0077] Step S219, calculating the deviation value between the flow data and the reconstructed data;

[0078] Specifically, a difference operation is performed on the flow data and the reconstructed data to obtain a deviation value; the above deviation value can be set to res, then res = y-y', and the corresponding deviation value of the import flow rate is res1 = y1-y'1, and the deviation value of the import flow rate is res2 = y2-y'2. The deviation fault identification can be performed on the import or export flow rate separately, or on both at the same time, without specific limitation here.

[0079] Step S221: If the deviation value exceeds a preset deviation threshold, it is determined that a deviation fault occurs in the diesel engine of the ship.

[0080] Specifically, the above res1 and res2 can be compared with δ1 and δ2 respectively. If one or both of res1 and res2 exceed the corresponding δ1 and δ2, it can be determined that the diesel engine of the ship has a deviation fault.

[0081] In practical applications, the above method can be used to monitor the fuel inlet rate and outlet rate of the diesel engine flow meter in real time when the ship is sailing. When it is determined that any deviation occurs in the inlet rate and outlet rate of the ship's diesel engine, the corresponding fault alarm information of the fault location can be fed back to remind the staff to check the diesel engine.

[0082] Corresponding to the above method embodiment, the embodiment of the present invention provides a processing device for identifying abnormal fuel flow rate, Figure 4 A schematic diagram of the structure of a fuel flow abnormality identification device is shown. Figure 4 As shown, the fuel flow abnormality identification device includes:

[0083] The acquisition module 401 is used to acquire the flow data of the ship underway; wherein the flow data includes the inlet flow rate and outlet flow rate of the diesel engine fuel of the ship;

[0084] The training module 402 is used to input the flow data into a pre-trained deviation fault recognition model, reconstruct the flow data through the deviation fault recognition model, and output the reconstructed data corresponding to the flow data;

[0085] A calculation module 403 is used to calculate the deviation value between the flow data and the reconstructed data;

[0086] The determination module 404 is configured to determine that a deviation fault occurs in the diesel engine of the ship if the deviation value exceeds a preset deviation threshold.

[0087] The embodiment of the present invention further provides an electronic device, such as Figure 5 As shown, it is a schematic diagram of the structure of the electronic device, wherein the electronic device includes a processor 51 and a memory 52, the memory 52 stores machine executable instructions that can be executed by the processor 51, and the processor 51 executes the machine executable instructions to implement the above-mentioned method for identifying abnormal fuel flow.

[0088] exist Figure 5 In the illustrated embodiment, the electronic device further includes a bus 53 and a communication interface 54 , wherein the processor 51 , the communication interface 54 and the memory 52 are connected via the bus.

[0089] The memory 52 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 54 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used. The bus may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0090] The processor 51 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 51. The above processor 51 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor to execute, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor 51 reads the information in the memory 52 and completes the steps of the method for identifying abnormal fuel flow in the above-mentioned embodiment in combination with its hardware.

[0091] An embodiment of the present invention further provides a machine-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the above-mentioned method for identifying abnormal fuel flow. The specific implementation can be found in the above-mentioned method embodiment, which will not be repeated here.

[0092] The computer program product of the method, device and electronic device for identifying fuel flow anomaly provided in the embodiments of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the method for identifying fuel flow anomaly described in the previous method embodiment. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0093] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0094] In addition, in the description of the embodiments of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0095] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0096] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for identifying abnormal fuel flow, characterized in that: The method comprises: Acquiring flow data of a ship underway; wherein the flow data includes an inlet flow rate and an outlet flow rate of diesel engine fuel of the ship; Inputting the flow data into a pre-trained deviation fault identification model, reconstructing the flow data through the deviation fault identification model, and outputting reconstructed data corresponding to the flow data; Calculating a deviation value between the flow data and the reconstructed data; If the deviation value exceeds a preset deviation threshold, it is determined that a deviation fault occurs in the diesel engine of the ship; The deviation fault identification model includes an encoder, a hidden layer representation layer and a decoder connected in sequence; The method further comprises: Initializing a pre-established deviation fault identification model; Obtain pre-constructed training samples and train the deviation fault recognition model according to the following iterative process until a preset training stop condition is met: Input the training sample into the encoder of the initialized deviation fault recognition model, and map the training sample to the latent space corresponding to the hidden layer representation layer; reconstruct the training sample through the decoder, and output the reconstructed sample corresponding to the training sample; calculate the loss function of the training sample and the reconstructed sample; The training stop condition includes at least one of the following conditions: the number of iterations reaches a preset number, or the loss function converges.

2. The method for identifying abnormal fuel flow according to claim 1, characterized in that: The method further comprises: Acquire historical flow data of the ship during navigation, wherein the historical flow data includes a historical fuel inlet flow rate and a historical fuel outlet flow rate of a diesel engine of the ship; The historical traffic data is preprocessed to obtain the training samples, wherein the preprocessing process at least includes data cleaning and data normalization.

3. The method for identifying abnormal fuel flow according to claim 1, characterized in that: The method further comprises: Calculating residual values ​​of the training sample and the reconstructed sample; A kernel density analysis is performed on the residual values ​​to find a deviation threshold for characterizing the normality and difference of the training samples.

4. The method for identifying abnormal fuel flow according to claim 1, characterized in that: The method further comprises: The training samples are mapped to the potential space corresponding to the hidden layer representation layer using the following formula: h=f(wx+b) Among them, f(wx+b) is the coding transfer function, x is the training sample, w is the coding weight of the coding transfer function, h is the eigenvalue of the hidden layer representation layer, and b is the bias parameter of the coding transfer function.

5. The method for identifying abnormal fuel flow according to claim 1, characterized in that: The method further comprises: The decoder reconstructs the training sample using the following formula and outputs the reconstructed sample corresponding to the training sample: x'=g(vh+b') Among them, g(vh+b') is the decoding transfer function, x' is the reconstructed sample, v is the decoding weight of the decoding transfer function, h is the eigenvalue of the hidden layer representation layer, and b' is the bias parameter of the decoding transfer function.

6. The method for identifying abnormal fuel flow according to claim 1, characterized in that: The loss function is expressed as the following formula: L=||x-x'|| 2 Wherein, x is the training sample, x' is the reconstructed sample, and L is the loss function.

7. A fuel flow abnormality identification device, characterized in that: The device comprises: An acquisition module, used to acquire flow data of a ship underway; wherein the flow data includes an inlet flow rate and an outlet flow rate of diesel engine fuel of the ship; A training module, used for inputting the flow data into a pre-trained deviation fault recognition model, reconstructing the flow data through the deviation fault recognition model, and outputting reconstructed data corresponding to the flow data; A calculation module, used for calculating a deviation value between the flow data and the reconstructed data; A determination module, configured to determine that a deviation fault occurs in the diesel engine of the ship if the deviation value exceeds a preset deviation threshold; The deviation fault identification model includes an encoder, a hidden layer representation layer and a decoder connected in sequence; The device is also used to: initialize a pre-established deviation fault identification model; obtain a pre-constructed training sample, and train the deviation fault identification model according to the following iterative process until a preset training stop condition is met: input the training sample to the encoder of the initialized deviation fault identification model, and map the training sample to the latent space corresponding to the hidden layer representation layer; reconstruct the training sample through the decoder, and output the reconstructed sample corresponding to the training sample; calculate the loss function of the training sample and the reconstructed sample; wherein the training stop condition includes at least one of the following conditions: the number of iterative processes reaches a preset number of times, or the loss function converges.

8. An electronic device, characterized in that: The invention comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the method according to any one of claims 1 to 6.

Images